CONTENT CREATION

On-demand restore with a low-confidence human review gate

Accepts a webhook request to restore a specific archive image, runs Replicate, and routes outputs by quality: clean results auto-publish to R2 while low-confidence ones create…

CategoryContent Creation
Enginesim
Difficultyadvanced
Triggerwebhook
Steps5
Setup~25 min

How it runs

The automated pipeline, trigger to output.

  • TriggerWebhook: restore image requestHTTP webhook
  • ActionFetch source and run Replicate restore modelReplicateReplicate
  • LogicBranch on restoration confidence score
  • ActionHigh confidence: upload to public R2CloudflareCloudflare R2
  • OutputLow confidence: stage in R2 and open Linear review issueLinearLinear

What it does

Exposes a webhook so any internal tool can request restoration of one archive image. The output is scored, and the pipeline branches: high-quality restorations publish straight to R2, while questionable ones are held for a human to review in Linear before going live.

When to use it

Use it when restoration quality varies and you cannot blindly publish — for example fragile or heavily damaged originals where a bad upscale would embarrass the brand. The review gate keeps a human in the loop only when it matters.

How it works

  1. 1A webhook call with the image reference and metadata triggers the run.
  2. 2The source image is fetched and sent to the Replicate upscale-and-restore model.
  3. 3A confidence branch evaluates the model's restoration score against a threshold.
  4. 4High-confidence outputs upload to R2 and the caller gets the live URL.
  5. 5Low-confidence outputs upload to a staging R2 prefix and open a Linear issue with before/after links for an editor to approve or reject.

Set it up

What you configure once, before turning it on.

  1. 1
    Connect HTTP webhookTrigger any URL on agent actions.
  2. 2
    Connect ReplicateImage, video, and model inference.
  3. 3
    Connect Cloudflare R2Object storage, S3-compatible.
  4. 4
    Connect LinearIssues, projects, cycles, triage.
  5. 5
    Set each agent's modelWe leave models unset so you pick the tier — fast + cheap, or top-quality.
  6. 6
    Tune it to your dataEdit the prompts, filters, and field mappings so it matches how your team works.
  7. 7
    Test, then turn it onRun once against a sample, confirm the output, then enable the trigger.

Run this workflow in your colony.

14-day trial. No DevOps. No Sales call. Provisioned in under a minute.